Arthur T Knackerbracket has found the following story:
The skills gap is widening between people and AI.
Artificial Intelligence is apparently ready to get to work. Over the next three years, as many as 120 million workers from the world's 12 largest economies may need to be retrained because of advances in artificial intelligence and intelligent automation, according to a study released Friday by IBM's Institute for Business Value. However, less than half of CEOs surveyed by IBM said they had the resources needed to close the skills gap brought on by these new technologies.
"Organizations are facing mounting concerns over the widening skills gap and tightened labor markets with the potential to impact their futures as well as worldwide economies," said Amy Wright, a managing partner for IBM Talent & Transformation, in a release. "Yet while executives recognize severity of the problem, half of those surveyed admit that they do not have any skills development strategies in place to address their largest gaps."
[...] IBM says companies should be able to close the skills gap needed for the "era of AI," but that this won't necessarily be easy. The company said global research shows the time it takes to close a skills gap through employee training has grown by more than 10 times in the last four years. That's due in part to new skills requirements rapidly emerging, while other skills become obsolete.
(Score: 5, Insightful) by Rupert Pupnick on Monday September 09 2019, @06:32PM (4 children)
IBM has been pushing AI-as-a-Service as their main offering in their sales and marketing campaign for a few years now. This just fits into their narrative of "use our AI or get left behind."
The problem is that this story doesn't make any sense in the context of IBM's other big message which is that AI is going to make life easy as well. With any expert system you need at your beck and call, why the hell would you need any special training to close a "skills gap"? The AI is supposed to provide the skills!
I don't believe anything these guys say.
(Score: 4, Interesting) by ikanreed on Monday September 09 2019, @07:11PM (1 child)
The big AI scam of the 80s was that the AI algorithms couldn't do anything useful at all. The famous "Your neural network can't learn xor" paper.
The big AI scam of the 2010s is that the AI algorithms can solve a couple of kinds of problems about as well as humans, as long as the problem itself is well defined and you know how to tell the AI when it's wrong.
For example, ResNet can, as of 2017, classify images into pretrained categories a little better than a human being, and thanks to category agnostic pre-trained "base" convolution layers to its network, learn with a much smaller dataset. That's huge, but it's not "make a decision, act on it, and be able to explain your reasons" that is our corporate dream of AI. And even further from our "cognitively similar to human reasoning" dream.
(Score: 1) by kramulous on Tuesday September 10 2019, @02:50AM
Thank you two for saying that. I've thought I was going crazy since I am yet to see AI/ML work in a single instance. The excuse has always been "We just need more training data". And the executive level lap that shit up.
(Score: 2) by PartTimeZombie on Monday September 09 2019, @08:57PM
Me either. It's IBM. They have no clue what they're doing.
(Score: 1) by khallow on Tuesday September 10 2019, @04:42AM